Europe has chosen a different path from that of the United States and China. To achieve this, it has promoted a policy focused on technological sovereignty, security, and the defense of democracy, with the aim of guaranteeing the technological resilience of the Member States and protecting the European model of welfare in the face of possible geopolitical conflicts.

This means that Europe must be able to manufacture its own chips and develop algorithms designed and trained under European standards. All of this without abandoning the ethical dimension: an AI Made in Europe aspires to be synonymous with trustworthy intelligence.

The European Artificial Intelligence Act

The EU Artificial Intelligence Act was the first comprehensive regulation in the world dedicated to this technology. It will fully enter into force in August 2026. From that moment on, any company operating in the EU will have to comply with strict rules of transparency and accountability.

As happened with the General Data Protection Regulation (GDPR), this legislation is having a global effect: many American and Chinese companies are adapting their models to comply with European regulations so as not to lose access to a market of more than 450 million people.

Leading European AI companies

Europe does not have its own “Google”, but it does have companies that compete in efficiency, specialization, and privacy:

Mistral AI (France): considered one of Europe’s major bets. Its models, such as Mistral Large 3, stand out for their efficiency and for the possibility of being installed “on-premise”, that is, on the company’s own servers, ensuring that data does not leave the organization.
Aleph Alpha (Germany): strongly oriented toward the business and governmental sectors. Its approach focuses on transparency: its systems can indicate precisely which document supports each response, something essential in fields such as justice or medicine.
DeepL (Germany): continues to outperform many competitors in professional translation thanks to an AI specialized exclusively in language.
ASML (Netherlands): is a key element of the global supply chain. Without its advanced lithography machines it would not be possible to manufacture next-generation chips anywhere in the world.

Infrastructure: AI factories and gigafactories

Europe has understood that it cannot depend exclusively on large American cloud providers such as Microsoft (Azure) or Amazon (AWS). For this reason, in 2018 it launched the EuroHPC Joint Undertaking initiative, aimed at developing European supercomputing.

Among its achievements are:

• The acquisition of high-performance supercomputers such as JUPITER (Germany), LUMI (Finland), Leonardo (Italy), and MareNostrum 5 (Spain).
• The creation of “AI gigafactories” that allow European startups to train models at low cost.
• The promotion of quantum computing, with the inauguration of the first EuroHPC quantum computers.

Another fundamental pillar is data sovereignty: European AI is encouraged to be trained with European data (books, historical archives, industrial data) in order to avoid external cultural biases.

Recently, in March 2026, EURO-3C was launched, an ambitious project to create a federated European network integrating telecommunications (5G and 6G), edge computing, and cloud services under European jurisdiction. Unlike centralized clouds, EURO-3C connects operators such as Telefónica, Deutsche Telekom, and Orange in a distributed network. AI coordinates in real time where data should be processed most efficiently, reducing energy consumption and external dependency.

🇪🇸 Artificial Intelligence in Spain 🇪🇸

Spain is one of the few European countries with two “AI Factories” approved by the European Commission.

The Barcelona Supercomputing Center (BSC-CNS) hosts one of the EU’s original factories and operates the MareNostrum 5 supercomputer, where sovereign European models are trained.

The CESGA has been selected to host 1HealthAI, specialized in life sciences and health.

Spain also has a national language model called ALIA, aimed at allowing public administrations and companies to work with AI trained in Spanish, Catalan, Galician, and Basque.

In addition, the company Multiverse Computing, based in San Sebastián, stands out in quantum computing applied to AI, developing technologies that make it possible to compress language models and drastically reduce energy consumption.

🇮🇹 Artificial Intelligence in Italy 🇮🇹

Italy has focused on an ecosystem centered on industrial data. The CINECA consortium in Bologna hosts the Leonardo supercomputer and the IT4LIA factory, specialized in advanced manufacturing and the automotive sector.

The company iGenius, based in Milan, has specialized in business intelligence. Its platform Crystal connects directly with corporate databases, reducing the risk of “hallucinations” typical of more generalist models. It is a Small Language Model (SLM) designed for maximum privacy and compliance with European regulations.

🇵🇹 Artificial Intelligence in Portugal 🇵🇹

Portugal has positioned itself as a gateway for data and a center for large-scale computing.

The Deucalion supercomputer, located at the Minho Advanced Computing Centre, supports both the public sector and local startups.

The country has proposed a project to build a gigafactory in Sines, taking advantage of submarine cable landings and seawater cooling for a more sustainable model.

Among its companies are:

• Feedzai: specialized in real-time financial fraud detection. It operates on a global scale and is a reference in explainable AI (XAI), aligned with European regulations.
• Defined.ai: focused on the creation and sale of ethical and legal datasets to train AI models, promoting multilingualism and avoiding practices of massive data extraction without consent.